Infrared scanner for sandalwood

Viewing an extensive area of any crop from the air can save labour and improve crop knowledge and the accuracy of ground level observations.

When you use near-infrared imagery, a whole new set of data becomes available, with amazing applications.

When Haefeli Lysnar business development manager Garth Johnson travelled to the Carnarvon research station to demonstrate the application of UAVs in horticulture, he received interest from an unexpected quarter.

His objective was to show the benefit of aerial photography in the management of traditional plantation crops, unaware that Agriculture WA had trial plots of sandalwood being assessed.

When he mentioned he had near-infrared images of the fledgling trees to a sandalwood company, they were extremely interested to see what information they could glean from the photographs.

The sandalwood growers had come to Haefeli Lysnar to find out about using a mobile laser scanner to measure growth of the sandalwood plants from the ground.

Such measurement was going to be unsuitable for their purpose, but they were delighted to see the amount of information that could be obtained from the aerial images, a method they had not considered.

International investors, who require regular reports on the current and projected value of their plantations, own the sandalwood project.

Manual ground surveys are expensive and inaccurate. A manual inspection of 10 per cent of the trees has an error rate of up to 30 per cent and can cost more than $1200 per 1000 hectares, and these are required frequently over the life of the trees.

Sandalwood is a root hemi-parasite and requires host trees, usually acacias, for healthy growth. The host trees provide extra water and nutrients to the sandalwood. The two trees form a symbiotic relationship throughout the life of the sandalwood tree.

"We flew a senseFly eBee over the trial plot, photographing near-infrared imagery from a height of 100 metres," Mr Johnson said.

"With normal red-green-blue (RGB) images we couldn't differentiate between the Acacias and the sandalwood growth, but with the NIR band the difference was dramatic.

"From the NIR images we then produced NDVI - normalised different vegetation index - a comparative index that shows a scale from negative 1 to plus 1. Negative 1 to 0 is non-vegetation, whereas 0 to plus 1 is vegetation.

"When we looked look at how things were behaving from 0 to plus 1, the images showed a very distinctive variation in colour between the host tree and the parasitic sandalwood."

While the original company inquiry for a more accurate method of assessing the state of the sandalwood crop was not practicable with ground surveying technology, the fly-over information solved an even larger problem.

The image clearly differentiated the sandalwood from the non-commercial host tree. Sophisticated software then counts the numbers of sandalwood plants and calculates the density per hectare.

Not all plants that are introduced to the host plant will take, so an accurate picture of the successful plants is also determined.

From the density per hectare, the estimated value of the entire plot can be determined. This is a very time-consuming and expensive operation if undertaken physically on the ground.

Aerial NIR images taken throughout the five-year growth to maturity of the sandalwood can provide other important data such as the mortality rate, rate of growth, management strategies and further value estimates.

An additional benefit that excited the sandalwood producers was the ability to determine the size of the canopy of the crop.

"Knowing the size of the canopy will give the producers information that they can use in the future to correlate the amount of wood yield with the size of the canopy covered by the sandalwood," Mr Johnson said. "This gives them a far more accurate estimation of the amount of commercial sandalwood they can expect. Propriety software will allow growers to determine this information automatically."

Another unexpected advantage of the NDVI imagery was where a sandalwood bush that would have been obscured by an overgrown host tree, and therefore missed in a RGB photograph, could be easily differentiated.

"The images collected from an UAV can give a far more accurate assessment and an improved information base and will pay for itself in the first few surveys. It takes so much of the human cost and error out of the audit," Mr Johnson said.

"While we knew of the application of this technology in horticulture, viticulture and broad acre agriculture, we can now see further use in specialised crops like sandalwood."